In addition to the next-generation GeForce RTX 4090 and RTX 4080 gaming graphics cards based on the Ada Lovelace architecture, NVIDIA introduced a professional graphics accelerator for NVIDIA RTX 6000 workstations based on the same architecture at the GTC conference. The manufacturer notes that the new product offers 2 to 4 times higher performance in professional tasks compared to the previous generation RTX A6000 model based on the Ampere architecture.
The main features of the Ada Lovelace graphics architecture, and therefore the graphics cards based on it, are: third generation RT cores, which provide up to 2 times higher performance in ray tracing operations compared to the previous generation; 4th generation tensor cores that are up to 2x faster in AI learning, with improved support for the FP8 data format; and also CUDA cores, up to twice as productive in single-precision operations compared to previous generation solutions.
The professional graphics accelerator NVIDIA RTX 6000 received 48 GB of GDDR6 memory with error correction function (ECC). The specified power consumption is 300 watts. The novelty is equipped with an active cooling system and works with a PCIe 4.0 x16 interface. The card offers four DisplayPort 1.4 connectors.
The professional card also claims support for vGPU virtualization technology and up to 3x faster video encoding performance for streaming multiple simultaneous XR sessions with NVIDIA CloudXR compared to the previous generation.
The NVIDIA RTX 6000 professional graphics card will be available in December from the company’s main partners. The price has not yet been announced.